« Open Thread 2 | Main | Open Thread 4 »


Feed You can follow this conversation by subscribing to the comment feed for this post.


Here's FrankD's post in the last Open Thread.


I've been playing with PIOMAS figures lately. I note the PIOMAS anomaly graph has a linear trend superimposed, but the data is pretty obviously not linear. The maximum anomaly does not correspond with the minimum volume, so its hard to read for info on when we will reach "zero ice".

A linear trend has a R-squared of around 0.85, where a 2nd order polynomial is .925. Probably, an s-curve would be better, but I don't know how to do them.

Anyway, I graphed the monthly figures: average of all anomly data for the month + mid-month daily average (apples and oranges, I know, but it suited my purpose).

What comes out? A better than 50% probability of an ice free September by 2016, with the ice free period increasing by almost a month each year. By 2023, there is a good likelihood of five months ice free, from mid-July to mid December. After that it slows down somewhat, but March and April, the last months remaining, reach zero around 2032-33.

I've posted the results here:

I'd appreciate any thoughts on how to improve this - I have a good body of data behind it, its really just a question of the best way to slice and dice it. As I say, a 2nd order poly gives artificially quick zeroes, but since a lot of people mean less that 10% of the long term average when they say "zero", it kind of balances out.

Andrew Xnn

2033 - 2016 = 17 years... Wow!

So, there is data supporting a tipping point.

I'd like to validate the math, but after poking around the PIMOS site have not
been able to find the data set. If you could point me in the right
direction, I'll put it into Excel and check it out as well.

Curve fitting data has its limitations since the actual melting is composed of a myriad of physical processes, feedbacks and other things. Still an excellent exercise and perspective.


An interesting read concerning the NAO and the AO and the influence they are currently heaving on the Northern Hemisphere. Via RC comments.


Neven, I suspect we might be coming out of hibernation early. The sea ice is doing strange things and I've also noticed that the AO forecast and even the ENSO forecast seem to be off track. And, as we've all noticed, the weather is really strange! Also the new Icesat will be available shortly, which will be the start of the ice season.


Meanwhile we have a biologist saying in the New York Times:

The notion that no "tipping point" exists for Arctic ice decline has spread in climate science for several years, supported by deeper examination of the North's physics. Initially, the media exaggerated fears that the loss of ice, which naturally reflects light, would expose more heat-absorbing water to the sun, causing runaway decline. However, scientists now widely believe this feedback is balanced by a host of other phenomena, like increased flows of hot air from the tropics, improved ice formation efficiency under thinning conditions and the region's general cloudiness.

The article is filled with things that go against practically everything I've learned since I started this blog. So either my alarmism prevented me from taking a broad view of the literature and thus playing down the uncertainties in my little confirmation bias-world, or something is really wrong with one NYT reporter and a couple of scientists.

Here's some more stuff:

Those fears, which peaked in 2007, were likely exacerbated by the stark retreat in sea ice that year. The Arctic lost more than 1.6 million square kilometers of ice, an area larger than Alaska;

1.6 million square kilometers of ice? Are you kidding me?

However, since that shocking decline, the ice has modestly expanded during the summer, perhaps the best evidence that Arctic ice won't drop off a cliff, said Eric DeWeaver, a co-author on the Nature paper and physical climatologist at the National Science Foundation.

The 2007 loss was "spectacular," he said, but "one would not expect to see it very often."

Where was this guy this last melting season?

It's uncertain what physical process helps stabilize the warming seen from exposed water. One important effect is that as ice declines in thickness, it becomes more efficient at growing, a well-established truism. Cold air at the ice's surface causes heat to rise through the slab, all the way down to the ice-water interface, where more ice then forms. The efficiency of this exchange improves as the slab thins out, helping floes winnowed by summer temperatures grow back in the winter.

Can someone confirm this?

Also, there are simple facts like the persistent clouds that ring the Arctic and lack of sunlight during the ice's natural minimum that could neutralize declining reflectivity.

Really, is that a simple fact?

Modeling sea ice decline is notoriously difficult, and there remains wide disagreement in the models on how quickly the Arctic will retreat under warmer temperatures. Several models project a complete loss of summertime ice before the century's end, while others chart a modest 15 percent decline.

FYI, this article was written a few days ago, not in 1997!

I'm in a bad mood, and this article hasn't improved it...


Thanks Arossam,

Bit of edit-fail on my post - I tried several times to post that and it kept rejecting. I thought it was length, then links, then funny characters, but it was just a random glitch. Anyway, in the succession of edits to try to get it to take, I lost a key point - Polar Science Centre do not publish a data set at this time. I constructed it off their graphs.

For each horizontal point I could resolve from their graph (each point about 9.3 days apart) I took the best mid point I could determine. I then averaged that over the month (3 or 4 points per month, weighted - so a point that covered, say, 25 Feb to 6 Mar would be allocated 1/3 to Feb, 2/3 to Mar).

That gave me a linear data set that replicated the PIOMAS graph pretty well. See:
and compare to the original:

I've added the linear and poly trends with formulae and R2 values. X-axis check marks are at 1 year intervals; the number there is just the number of data points @ about 39 per year (157 every 4 years to be exact).

While interesting, that didn't achieve what I wanted. The maximum negative anomaly occurred in June 2010, and doesn't correspond to the minimum actual volume. I was curious about predicting when we would see the volume hit zero (which I expected would be in September), so I needed data for other months. So I broke the dataset down into monthly totals, and piped it all through Excels chart functions. As noted adding averaged monthly anomalies to their mid-month average is a bit dodge, but you work with what you have.

But I completely agree with your last point - trend projection only works when your method of extrapolation mirrors some physical reality. My choice of 2nd order poly was simply that trusty old Eyeball Mk I said that would be a better fit than linear. As I said before, I believe the actual decline will follow an S-curve, which is beyond my poor powers to overlay on the data I have.



I think your NYT commentator makes some good pointstouching on some negative feedbacks that emerge only after a certain amount of ice is gone.

I don't know if his description of heat transport through ice is correct, but *loss* of ice will help transport heat from the ocean to the atmosphere to space (since there's no ice/snow to insulate it), so as we lose more extent, the rate of loss will slow, all other things being equal.
Warm ocean (without ice cover) + cold air should create more fog / cloud, so again, more expanse of open water should lead to less total insolation at the surface (I think the weird weather we saw this year might become rather typical). I do expect cloud cover to increase as ice extent/area decreases.

But I think he also misses some biggies:
1. A lot of heat transport is coming via the ocean, not directly from the sun. Cloud over the Arctic is less important if warmer water from temperate regions is doing an A-grade job of melting ice.
2. I dont think there is a tipping point in the future. I think we've already passed it. Depending on whose numbers you are using, and whether you are looking at area or volume or duration of ice coverage in different places, somewhere between 2002 and 2006 we took a downward turn. The 2007 season wasn't something bad happening, it was a symptom that something bad had *already* happened.
3. 15% by 2100. Laugh? I nearly started... PIOMAS modelled a 25% loss in September volume between 2009 and 2010. Extent is holding up better than volume, as the ice spreads out, but that will collapse when volume gets low enough. Does anyone really think 3 million sq kms of 1 mm thick ice is a physical possibility?

What I get out of it is a great cynical LOL at the spintastic sub-editing:
Headline: No Tipping point for polar bears.
Text: "...hope that, with aggressive greenhouse gas emissions cuts and wildlife management, polar bears may retain viable habitat into the next century."
Lets just focus on that: hope..aggressive cuts..management..may retain..habitat..next century.
Subtext: If we could do things that we won't actually do, we hope we might manage to keep polar bears going in a few areas for another century, before they die out. But probably not.

What was the phrase? The "audacity of hope"? After that heady dose of optimism, I need a lie down.....


Neven: You hit the nail on the head on your post @ "In it for the Gold" ( see my comment there). I encourage commenters on this blog who are inetrested in the "big picture" to have a read of Neven's post.


That's awfully nice of you to say, Phil, thanks a lot. I'm just beginning to scratch the surface of the tip of the iceberg, but I think I've got the right iceberg. :)

Thanks for commenting on that NYT piece, Frank. I had to blink several times with my eyes while reading it.

Peter Ellis

The bit about heat transport through ice is trivially correct. Ice is an insulator, and once the ice cover has formed, heat loss slows dramatically. The ice thickens at an ever-decreasing rate until the insulation is sufficient to prevent further growth - a couple of metres thick. Growth beyond this comparatively thin skim takes many years of compression / ridging / rafting.

In practice, what this means is that in the Arctic winter, first year ice will always form and grow to a couple of metres thick. Loss of the winter ice seems extremely unlikely. After all, if winter ice forms on Hudson Bay, at lower latitudes (and it does, every year), why would it not form further north? The maximum winter ice extent is unlikely to drop too dramatically, as however you cut it you will still be able to form a thin skim of a couple of metres of first-year ice across the Arctic Basin in the depth of the Arctic winter. However, that argument, since it focuses on extent rather than volume, misses the wider point that the Arctic is not (yet) a seasonal pack, and the remaining old ice up there is much more than a single year's thickness. Volume loss in winter is ongoing, as the PIOMAS anomaly graph shows: and the thinner the winter ice cover, the easier it will melt out in the summer.

The bit about albedo is again trivially true. Broadly speaking, the thaw starts as the Sun rises above the horizon, and continues until it drops below the horizon. Thus, the exact minimum will necessarily occur just at the point the Sun sets - so the albedo effect of a worse (lower) minimum won't be large. However, again this neglects the important point: a lower minimum means lower ice levels leading up to the minimum, i.e. when the sun is high in the sky and the albedo effect significant. What matters most for the ice/albedo feedback is June ice levels, not September ice levels - yet a low September level almost certainly means that the June level was also low.

Andrew Xnn

Thanks for the explanation FrankD; an impressive feat to convert their graph into usable data! Extrapolating historical data has both limits and value. I'll guess that it does very well over the short term, but has limited value beyond 5 years or so.

Physics based models on the other hand are nearly impossible to build. There are too many factors that need to be taken into account and so far the best in the world are missing the mark.

I agree with Peter Ellis about these physical processes, but instead of describing them as trivial, I'd warn readers to not focus on only one or two mechanisms. Unless one can quantify all the many feedbacks, then speculating about how they may influence the longer term behavior is likely to result in gross errors.

The NY Times article had a number of links to good papers. Eisenman for example points out how the placement of the continents has caused seasonal asymmetries with the melting. He finds that zonal-mean sea ice edge latitude is easier to predict and advancing at the rate of 8 km/year. Of course, the thickest sea ice is a long way from the geometric north pole.

Winton points out that while some sea ice models have problems which can not be excused, substantial natural variability is needed to explain why the others are missing the mark.

Anyhow, over the short term we can expect about 350 km^3 of arctic sea ice melt per year. This rate may be accelerating, but it is too soon to know for sure.


Neven, "I'm in a bad mood, and this article hasn't improved it..."
Doesn't the phrase "bear with a sore head" relate to their behaviour when they get woken during hibernation?!

I presume you have seen this by now, he sounds annoyed too:



Hey Lodger! Cryosat delivers OK! This is amazing to actually see the Beaufort Gyre, I can hardly wait for more.


Arossam: "Unless one can quantify all the many feedbacks, then speculating about how they may influence the longer term behavior is likely to result in gross errors."
Yes, but we amateurs can have a lot of fun being spectacularly wrong along the way. Man proposes, nature disposes...
But I would quibble on one point you make: "Anyhow, over the short term we can expect about 350 km^3 of arctic sea ice melt per year. This rate may be accelerating, but it is too soon to know for sure."
This was *exactly* my point - it is very clearly accelerating. I can't state that it will continue to do so, but using the better-fitting quadratic on that second pic I posted above gives a current slope not of 350 km^3 (the "comforting" linear slope from PIOMAS) but 850 km^3. Someone may want to check my calculus on that - the derivative of the formula on the second graph.

For what that's worth, 850 per year x 4 years = ice free September.

Peter - very nice explanations. I would just point out that James Bay has ice cover in the low 50's North while there is still open water near Svalbard in the high 70's North, so its not just latitude. There is a warmish current flowing from the South into the Barents Sea, while Hudson Bay, if it gets any inflow, is cold water from the North. I also suspect that Hudson Bay, being relatively shallow and enclosed is probably more likely to cool more quickly (partly due to stratification, partly due to lower surface area / volume ratio).

Neven - congrats on the post at mt. Good stuff, but you missed the chance in amongst those Churchill quotes to use his best (IMO):
"The era of procrastination, of half-measures, of soothing and baffling expedients, of delays, is coming to its close. In its place we are entering a period of consequences". November 12, 1936.
(hat tip to Al Gore)

Kevin McKinney

It's coming as much from my heart as my head, Frank, but WRT your Churchill quote--what a wordsmith the man was!--I can only say, "Just so."

Artful Dodger

Neven, Mary Timus may have a satisfying answer to your question in a comment posted on Climate Progress.

FrankD, good work mate.

If you haven't seen these yet, look at:

Arctic death spiral:” Naval Postgrad School’s Maslowski projects ice-free fall by 2016 (+/- 3 yrs)

Sea Ice Volume graphic on Wikipedia

My Oct 12, 2010 prediction made here on ASI that arctic sea ice could be gone by Sep 2014, and the first sea ice-free Winter could occur 9.5 years later, in Mar 2025 (more on this tomorrow).


Thanks, Lodger. I hadn't read those comments yet. Here's the homerun:

Amstrup is no longer with USGS but is now Chief Scientist with Polar Bears International. The latter organization needs to keep potential donors and collaborators optimistic. Take a look at the extent of their “Donate and Help” page on their website. They cannot admit they are essentially on a deathwatch given current and projected emissions – and the demonstrated inability of governments to curb those emissions.

FrankD, that is a great quote indeed. We are entering the period of consequences...


BTW, IJIS has ended its stint of extent decrease, 4 days in total. Quite exceptional at this time, but the graph clearly shows that.

Andrew Xnn


You're probably correct about the acceleration. Over the last 5 years we've witnessed much more than 350km^3/year ice loss and it's very difficult to accept the professionals position that this might be all due to natural variability as opposed to an actual acceleration and thus a gross defect in all the models. Of course, some people are more likely to accept the variability suggestion than others.

However, there is a fair amount of noise in the PIOMAS volume data; probably a 1000 km^3 at least. So, last years 4000 km^3 could be between 3 to 5000 and at 850km^3/year that would work out to somewhere between 4 to 6 years.


Arossam - you are quite right - the range of values would indeed give 4 - 6 years, not the four I suggested. I'm going to tinker with error bars when I get some downtime to try to tighten it a bit, but current numbers point to September ice reaching trivial levels in the 2014 - 2016 window. Extent analysis paints a much rosier (but I believe false) picture.

Cheers Lodger - yes Maslowski looms large on my radar. I recall when his paper first came out, I was a bit dubious, but thought on it some more when I saw the derision that greeted it on the other side of the mirror - what is important there is unimportant in reality (and vice versa) is a pretty fair rule for LookingGlass Land. I now consider Maslowski's 2016 centreline to be a very good bet. I'd forgotten your October call - 2014 might be a bit quick, but either way, we don't have that long to wait to find out. In any case, multiple different approaches converging on roughly the same conclusion...

The thing I note from Maslowski's data compared to the PIOMAS model results is how much less variable his early volumes are - about 2500 km^3 higher minima and 2500 km^3 lower maxima. The datasets converge from around 2002. In a physical sense, I'm not sure what the import of that is. It makes me wonder if the midmonth numbers for PIOMAS aren't a bit misleading. If the reality behind those averages is that the seasonal volatility is growing (which is not obvious from the anomaly data alone) that does suggest an earlier September zero, but a later April zero.

Mathematically, the more-rapid-than-average fall in minima in the period he is using for his linear trends (1995-2005) explains why his linear projection is so much more agressive than the PIOMAS one (calculating loss at >1000 km^3 per year). A linear trend through September PIOMAS data points to zero at about 2030.

Reducing the seasonal variation has the effect on my polynomial trend of pushing the September zero a few years later, but pulling the April zero back to the late 2020's. OTOH, If I just truncate the data and build my trends off that same 1995-> range, polynomial trends give a September zero in 2013 and perennial ice free by around 2023.

Generally his dataset (particularly the steep fall in the early noughties that brings the two datasets into closer alignment) points to a much greater "tippiness"in the whole situation, making the biologist from the NYT seem even more Pollyanna-ish

Probably wont be back before the 25th, so to all in Nevenland, a happy Christmas, Hannukah, Saturnalia or whatever other midwinter feast you observe.
"And a Festivus for the rest of us". ;-)

Peter Ellis

If seasonal volatility is growing, then we should start to see a yearly cycle within the anomaly data, with consistently greater anomalies during the summer melt and lesser anomalies during the winter/spring. This is very apparent for area at Cryosphere today, and there's a hint of it from 2007 onwards in the PIOMAS anomaly graph. Nothing like as big a difference as there is in Maslowski's graph though. I wonder what model he used for his ice volume figures? The (in development) PIPS 3.0, or something else?

Andrew Xnn

Comparing the two charts (area and volume), it is apparent that there has been a significant shift in the dynamics of sea ice since 2007. However, there has to be more uncertainty in volume measurement than in area.

Anyhow, this seasonality is probably all due to the ice becoming so much thinner that it can now spread out easier during the winter (or compact as we witnessed last week) depending on how the winds and current flow and blow.

This shift in dynamics has also got to be making the professional ice sheet modeler job much more challenging. I'll speculate that there is not enough data on this aspect for anybody to accurately model.



Blimey that is one hell of a jump.

Artful Dodger

Peter: See "Arctic Sea Ice Thickness Distribution: Modeling and Observations", Wieslaw Maslowski, Feb 16, 2010 (slides from his presentation to the Polar Climate Working Group).

In my reading on Sea Ice modeling, I've seen predictions that maximum variability will occur when SIE nears 3 M km^2. I'll see if I can find a reference to the paper, and will post a link here.

dorlomin: Yes, that's quite a spike in 80N surface temps. The DA has flooded the pole with warm air. Half the area North of 80 N (the Fram Strait side) was between -5 to -10 C on Dec 23, 2010.

CFS Alert on the Northern tip of Ellesmere Island had a high temp of +1.6 C yesterday, and a low of -6.9 C. Humidity is hovering around 75%, with SW winds 50 kph gusting to 70. You can see how broken and mobile the pack ice is in this ASAR image. Nares Strait is in the centre of this image, and Alert is near the top, slightly right of centre.

Pete Dunkelberg

"I'll speculate that there is not enough data on this aspect for anybody to accurately model."

I'll drink a Christmas toast to that, but Cryostat is on it! Will we have enough data before the ice is all gone?

"Blimey that is one hell of a jump."
I agree but clicking on several years it seems that jumps happen.

Artful Dodger

Merry Christmas!

The Dipole Anomaly (DA) will strengthen over the next week, as winds from the Greenland High are reinforced by a Low transiting NE from Iceland to Svalbard.

Greenland DA Animation

Watch for the PIPS sea ice arrows in the Greenland Sea this week while you enjoy a Happy Holiday! ☺


Merry Christmas everyone!


Merry Christmas to you Neven and to everyone else on the blog. Strange Christmas weatherwise indeed: It looks like its is warmer in Nuuk and in Alert ( as we hear from Lodger) than it is in Paris. Meanwhile, in Australia we are having a very wet Christmas with the first ( Category one) cyclone of the season making landfall just South of Cairns at 5 am this morning!

Andrew Xnn

Merry Christmas everyone from the Great Lakes region.

Not sure how a Greenland block and DA may or may not be related. However, the jet stream has been emptying the arctic into the Eastern US for a few weeks now. Has made for great skiing.

Anyhow, found that 80N temp profile from dorlomin very interesting.
Notice how unusual highs occur only during the winter.
During the summer, highs seldom occur and mean temp is only about 1C above freezing.

Also not much in 2010 compared to previous years to suggest record sea ice melt. So, maybe it is a Maslowski suggest primarily driven by ocean currents.


Hi everybody.

This is my first post here. I saw FrankD's graphics with some awe and went out replicating to see if it is really true.

Actually it could be worse, with possibly 2011 the last year with an all-year ice cover on the North pole.

Here is the september ice volume:


The data is surely not linear over the 1979-2010 period, so I fitted a straight line from 1995 to 2010. This predicts an ice-free september in 2019.
A quadratic fit (for the whole time span) to the data predicts 2016.
The best fit, available in Gnumeric, is a log fit and gives ice-free condition in 2012.

Of course these mathematical models do not mean all that much, yet I am confident that the real first ice-free September will fall in the 2012-2019 time frame. With ice-free probably meaning mostly ice-free.

I will be moving these spread sheet data to R now, to fit with some more physical models.

BTW, here are August and Oktober graphs:



I was looking at Topaz, and noticed just how much less 3m ice there is this year than last. However, the PIOMAS graph for December 2009 and December 2010 looked fairly close, though (despite the extreme dip earlier this year). Any comments? In particular, would any of the pixel-counters out there care to verify my intuition that PIOMAS is more optimistic than Topaz?





Welcome, wipneus! Great first post and love the name. ;-)

Great comparison, Bfraser. 2010 looks weird compared to 2009, perhaps due to the negative AO pushing things around? The discrepancy between PIOMAS numbers is interesting also. The last PIOMAS update was 3 weeks ago. The difference between the 2009 and 2010 Topaz maps around December 6th is even greater (if my eyeballs don't deceive me).

Here are the PIPS ice thickness images:


That seems to correspond better with PIOMAS.

Pete Dunkelberg

Wipneus: I will be moving these spread sheet data to R now, to fit with some more physical models.

Good idea. Extrapolating a curve fit into a confusion of processes is not exactly physics. But which physical processes are key? Irina Mahlstein's dissertation [Improving climate model projections by model evaluation and regional aggregation]
is of interest. Papers recently linked here discuss the open-ocean-lets-heat-escape-and-more-ice-form feedback idea. And there is the venerable "it's the wind blowing ice away" idea. Then there are clouds. When the summer wind dies down the ice can't blow away, and it also gets less sunlight (without wind,clouds are likely to build up). On the other hand winter clouds hold heat during the long night. Finally, the Hudson Bay area is too hot to believe.

You're going to be busy ;)


Andrew Xnn: The temperature remains ~1 degree C, but heat continues to build up. It just doesn't manifest itself as a temerpature rise - rather, the extra heat is consumed in melting ice. You would only expect to sea temps rise above that level once there is no more ice to melt.

Wipneus: very nice - this was exactly what I was hoping for (people with other stats tools fitting other curves, not an even earlier melt-out...) . I don't know which of these best represents the physical reality, but any of the curves would seem to be a better representation than a linear fit.
"Continue the research!" ;-)

Pete Dunkleberg: re:Hudson Bay - do you mean that you can't believe the temperatures in the same way that "I can't believe its not butter", or do you actually mean you think there is something wrong with the temp data? The multiple confirmations that southern and eastern Hudson Bay has no ice coverage confirms that the temps are definitely high.
I wonder though if the complexities you (correctly) point out don't fit into the "difference-that-makes-no-difference" basket. Even if the curve-fitting is inaccurate, on some level it does include the feedbacks you mention. A graph of global temps includes the result of increased CO2 forcing, but also negative solar forcing, aerosols etc, because temperature is the derived result of all these factors. In the same way, charting ice volume includes, to some extent, the results of the feedbacks you mention in the derived figure (volume). To get an exact timeframe, more accurate modelling would be required. But the timeframes seem to be so short, I don't see how the feedbacks can change the September result massively. I can see more impact on perennial results, but not seasonal. At least, they don't seem to have had much impact so far, outside of relatively quick autumn-winter growth, and we're not even seeing that this year.

Andrew Xnn


I appreciate that the temperature of an ice water mixture will not be much above freezing while it is melting and this is why summer time temperatures tend to hug 273K. However, in 2010 temperature 80N were actually a little lower than normal. Also, the melt season length was not remarkable. When I compare the earlier years 1958 or so to the last few years, there is no obvious trend (at least to my eyes).

Here is 1958:

Here is 2010:

Not much difference....

Andrew Xnn

Generally, not a good idea to try to answer your own questions, but here is how I would:

Maybe what is happening is that 80N temperatures are more a reflection of wind speed than what would normally consider.

During the earlier time periods (1958 or so), winds during the melt season were minimal and temperatures followed a similar and very predictable pattern. However lately, there is more wind and more heat being transported as cooler air from the ice is being mixed and transported southward.

During the melt season, temperatures 80N won't rise significantly above historical norms until there is a significant reduction in ice cover.

Link to Danish Meteorological Institute and Centre for Ocean and Atmosphere.




Hi Everyone, we live in interesting times. Hi wipneus, I'd like to see your graph with the log curve! My experience from project estimation is that many natural processes tend to follow log curves, it's the curve of growth but maybe the curve of decline as well. I've been looking at IJIS and something appears to be happening there. I was looking at Andrews graph of DMI and I felt I wanted to see the SD and anomaly for those temperatures going through the season and the area under the graph for the anomaly. I like that topaz stuff too, but I can't make much sense of it.

Artful Dodger

Pete said: "the Hudson Bay area is too hot to believe"

Believe it, it's called "Arctic Amplification", and it was one of the earliest predictions of the effects of increased CO2.

Here's a blink animation comparing Hudson's Bay / Baffin Bay on Dec 27, 2009 vs same date in 2010:

Hudson's Bay  Sea Ice Concentration Map 27 Dec


That's a h ell of a lot of open water, Lodger.

Any numbers for just that area?

Artful Dodger

Hi Adelady

Hudson Bay Sea Ice Area (including Foxe Basin, Hudson Strait and Ungava Bay) anomaly is currently -500,000 km^2. This is 45% of the Area anomaly for the entire Arctic (-1.12 M km^2).

Neven, you might consider adjusting the links on your Arctic sea ice graphs to include Hudson's Bay, Baffin/Newfoundland Bay, the Greenland Sea, and the Bering Sea, since that is where the action is this time of year. All 6 currently displayed Basins are near their seasonal maximums, and will remain so until May or June.

I hope everyone is safe and dry in Queensland.


Good idea, Lodger. I've added Hudson, Baffin and Bering (Greenland has also more or less reached its maximum and would screw up layout).

Great blink comparison, BTW.


Lodger. Safe and dry here on the Gold Coast and the Brisbane area generally. Rainfall for the month of December exceeded 550 mm ( as per my rain gauge, and consistent with BOM reports!). As we are close the coast, water quickly flushes out to the Pacific Ocean. However, areas in Central Queensland that received the same amount of rain were not so fortunate. They are now experincing once in a lifetime floods. In one particular town , all 300 + residents + pets had to be evacuated.
This is La Nina in Eastern Australia, and it is expected to be more of same until March.


Tamino has a good and informative post up: The History of Arctic Sea Ice Part 2.


Just a note on the little Petermann ice island, it's heading south! http://www.sailwx.info/shiptrack/shipposition.phtml?call=47557


"I hope everyone is safe and dry in Queensland."
Rumour has it all of Australia is in some kind of mourning. I believe they have lost a small urn they treasure greatly.


Kevin McKinney

". . . lost a small urn. . ."

Don't know too much about this, as a Canadian residing in the US, but I'm reminded of Twain's famous line about reports of his death having been "considerably exaggerated."

Evidently the same was true of English cricket--and will soon enough prove so with reference to Australian cricket, too. (Though maybe not "soon enough" for the fans.) If there's an area that outdoes climatology in the sheer prevalence of long-term predictions made on the basis of short-term observations, it's got to be sports.


Talk by Lonnie Thompsons wife on ice and ice cores



I've seen that 80N temperature graphic around, but I've not looked at it closely, so I can't really address your points in detail. However, if we agree that the latent heat of ice melting will hold the temps at close to zero, I'm not sure that you can determine the length of the melting season from temperature. And in any case I'm not sure that the length of the melt is particularly relevant - timing and intensity seem to be the keys.

You have given us one view of the melt. Here is another:

I've only looked at Hudson Bay so far. I reverse-engineered the standard CT graphs to add in the average, by adding the anomaly to the daily data. The delta between average and 2010 data is coloured in lime. What do we see? The duration of the melt was standard, perhaps a fraction longer. But it occurred on average about three weeks earlier than usual. The freeze up has been quite irregular against the average, due to unsually warm conditions in NE Canada, but overall is about three weeks later than usual. The rate of freeze, on average, is also slightly lower. Assuming that Hudson Bay does eventually ice over completely, a slower than average freeze up will mean that the freeze will be longer, but thats not a good thing (hence my question as to whether the length of the thaw / freeze is a good metric.

Overall, just for Hudson Bay, the length of the total thaw period (ie between the two seasonal crossings of any given area line) was some six weeks longer than average this year. The period for which HB was essentially ice free (<100K) was nearly eight weeks longer than average. That's extremely bad news for polar bears in that area (especially new mothers) - they had to get off the ice earlier, and many still haven't been able to get back out on it. There's an incredibly depressing video on this topic doing the rounds at the mo.

Hudson Bay is the extreme case, but there are similar trends in some other regions. Laptev, Kara and Barents Seas definitely show the same pattern to a lesser degree (I'll post similar here when I get around to them). Other seas (the ones fringing Siberia) may also, but are less obvious.

Lord Soth

Being the last day of the month, I decided to drop in to see the state of the ice; and it does not look good.

It look likes IJIS will ring out the year with under 12M km^2 of ice, and Eastern Hudson Bay and Hudson Strait is still wide open.

We are now more than half way thru the darkness, and the melt season will start in a little over two months. What ever ice that forms will be thin first year ice, and I exepct the ice extent to repeat the dramatic loss we saw in 2010.

I do get tired of explaning the difference between weather and climate to people. After a bad snowstorm, everybody thinks global warming is over. They fail to realize that with more heat in the atmosphere, we are going to get more extreme weather.

I keep saying snowstorms is weather; ice shelves turning into ice cubes is climate change; but they just don't get it.

Anyways, I shouldn't get upset with climate deniers. The laws of physics are not susceptible to rhetoric.

Anyways, its time to pour harmfull liquids in my body to celebrate another revolution around the sun. Happy New Years !

Pete Dunkelberg

Me: "unbelievable" of course I don't mean that I literally don't believe it, but Arctic amplification was never predicted to be so extreme. It's always worse than you think, it seems.

Lord Soth: "I do get tired of explaining the difference between weather and climate to people. After a bad snowstorm, everybody thinks global warming is over. They fail to realize that with more heat in the atmosphere, we are going to get more extreme weather."

With more water vapor (=latent heat) in the air we get more precipitation, wet or dry, except where we get drought (expanding Hadley cells). Send those folks to Climate Progress.

"The laws of physics are not susceptible to rhetoric." They're not susceptible to hope either, but human plans are. And so I say Happy New Year with this note of hope for an icy refuge.


Thanks for the encouraging reactions on my first post.

Pete Dunkelberg : sorry that the physical models I have in mind are not at all as sophisticated as that. I am just looking at something better than the mathematical functions available in a spreadsheet program that exhibit severe non-physical behavior.

MikeAinOz: the logarithmic fit predicts infinite decline in 2013, not acceptable for a physical model. Likewise the 2nd degree polynomial regression is a parabola, that, extrapolated before 1979, gives ice-free conditions earlier this century.
I agree interesting things are about to be seen on the ice-extent graphs.

The graph in : http://img718.imageshack.us/img718/4402/piomastrnd1.png shows the different trends that I looked at. The linear, 2nd degree poly and the logarithmic trends are the same as before. I added two more.

The first model, the purple line in the graph, assumes that the ice cover is driven by global temperature with first order time-lag. There is no self-amplification in this model. In contrast the second model assumes that change in ice-cover is driven by the amount of ice itself. This leads to an ice cover that changes with an exponential rate.

Assuming global temperature constant to 1970 and rising linearly after that, the "first order lag" model has a solution in the form
(M=ice mass; t = time in years; A,B and C regression constants) Physical reasoning let me limit the constant C to a few years. In the graph the limit was set at 20 years.

The second model assumes ice cover rate of change proportional to the ice cover:

d(M)/dt = (M-A)/C with solutions M= A - B * exp(t/C)
(M=ice mass; t = time in years; A,B and C regression constants)

Inspecting the graph hows that the exponential (the second) model fits the data best of all considered models. Compared with the disappointing goodness of fit of the simple temperature driven model, I believe this is an indication of self amplification of the sea ice decline.

Residuals are shown in http://img209.imageshack.us/img209/9580/piomastrnd3.png The exponential model was also applied to months June through December. These are plotted in http://img262.imageshack.us/img262/2294/piomastrnd2.png .

My conclusion would be that from these data, we can expect ice-free September conditions in 2014 (range 2013-2016). August will be ice-free within the year after that. In general we can expect ice volume August, July and June to decline faster than in October, November and December.

Anyway, I think I said enough for one post. All except all the best wishes for the new year.


All the best for 2011, everyone! I hope we know a bit more after this year's melting season.

Gail Zawacki

"Arctic amplification was never predicted to be so extreme...'

I watched this movie with Peter Cox about global dimming (http://witsendnj.blogspot.com/2010/12/insidious-soup.html) and was kind of shocked because I had no idea that the amount of sunlight reaching earth's surface has been reduced by 10% because of pollution...

and so I wonder is it possible it is warming relatively faster in the Arctic because their is less particulate matter in the atmosphere to mask the heating as there is the closer you get to the equator?

I would really love to hear from someone who knows something about this, because it seems that if it weren't for the pollution, we would be much much hotter with consequently even more violent weather than we are experiencing now.


Andrew Xnn



These papers examine some of the many problems with Arctic Sea Ice Models.


"These results suggest that most state-of-the-art GCMs are simulating observationally consistent presentday ice cover because the model errors associated with simulated cloudiness are being compensated by tuning parameters such as the ice albedo. In other words, errors in parameter values are being introduced to the GCM sea ice components to compensate simulation errors in the atmospheric components."

Winton, who is very kind towards the models, finds in figure 2 that every model seriously underestimates the amount of heat transfer from the Tropics to the Poles.

We amateurs can curve fit the data, but projecting it out beyond about 5 years is risky business.

r w Langford

Hi All Happy New Year.
Thanks for all the informative posts. I have learned so much from all of you. Hope some of you have seen the Nova production on the Antarctic recently. An ice free antarctic has been demonstrated from geological cores. Fossil leaves and organic mosses that have never thawed have been found. Lots of information on historic sea temperatures and levels as well as CO2 levels show the driving forces of climate change. Temperatures shifts less than are predicted for this century have had dramatic effects in the antarctic. Watch for repeat broadcasts. Keep up the great posts. Wish I could get the politicians in Canada to read this blog.


Gail, I distinctly remember seeing that BBC documentary 5 years ago, as it made a big impression one me.

As for your question: I'm not sure if aerosols are a large factor in anomalous Arctic warming. You can read more about causes of Arctic Amplification on SkepticalScience.

Unfortunately there is still a lot of uncertainty regarding aerosols and to what extent they cause global dimming.

Gail Zawacki

Hi Neven, actually Skeptical Science is the first place I looked. What I found there seemed to be saying that global dimming had been fixed because sulphur had been largely cleaned up since the '70's - but the movie leaves a very different impression.

It's like people saying the Montreal protocol was such a success and then I keep finding studies saying the ozone hole isn't fixed after all...

Pete Dunkelberg

Andrew: "We amateurs can curve fit the data, but projecting it out beyond about 5 years is risky business."

Projecting at all may be risky. Recall the argument for an oscillating extent. But I say
"Curve fits today, physics tomorrow!"
We are off to a fine start with wipneus's curve fits.

Somehow I was able to order "Principles of Planetary Climate" from Germany. It shipped yesterday (and the word "ship" may be literal) but it should soon be available everywhere. Unfortunately it does not cover ocean currents.

Daniel Bailey

Re: Gail

If you're Gail from Wit's End, I love your blog...

Following response reposted from over at Skeptical Science to an identical question there:

Cox is spot-on in regards to the protective effects (thus far) of the global dimming the video documents. This post covers that fairly well. Thus, as we clean up our emissions and therefore the air, we restore more fully the true heating capability of the sun, masked to some degree for decades. Add in a wakening sun (coming out of its dormancy as it climbs to the next solar maximum...

As to your last question, no. Arctic amplification of the high latitudes is the primary driver here. Romm over at Climate Progress touches on it in posts here, here and here as well (see also this related post). Basically, the warming of the Arctic, shifting of the polar jet and the expansion of the Hadley cells has reached the point where the entire circulation patterns of the northern hemisphere are being reorganized.

Hope this helps,

The Yooper


BTW, I hope one of our Japanese friends over at JAXA comes into the office tomorrow and updates the extent graph. It's been stuck at Dec 28th for far too long now.

Gail Zawacki

Thank you so much for reposting your explanation, Yooper - I had lost track of the Skeptical Science thread!

Daniel Bailey

It does get confusing sometimes over there. If there's a particular thread or question you're following, bookmark it or the last comment on it so you can find it again later. If John has you signed up as a member then you can use the member link to find your most recent comment.

Happy New Year to all!

(Had a question for Dodger & now I've forgotten...sucks to get old...)

The Yooper

Artful Dodger

Happy New Year everybody!

Welcome Wipneus and Gail, and a comment about IJIS:

Wipneus. Good work on the curve fitting. May I ask, where did you get the numerical data for sea ice volume? I haven't seen it posted anywhere. Your best fit curve produces good agreement with the estimates made by Wieslaw Maslowski (so no 'upturned nose' to that ;)

Gail, solar forcings in the Arctic are measured through remote sensing and direct field work. There was a large field experiment conducted during Summer 2010 by Scientists aboard the US Coastguard Icebreaker Healy in the Chukchi Sea. Their research goal was to measure the albedo of meltwater pools on the Arctic Sea Ice (a significant unknown for the Arctic Energy balance). One of the Scientists posts here if you care to search this blog. He also has his own science blog created during the mission on the Healy.

Modeling the atmosphere is more complicated. There are various differences near the pole, such as a thinner atmosphere (earth's rotation creates centrifugal force which forces air toward the equator), and an enhanced effect of black carbon on the albedo of sea ice. Nonetheless, models are validated against measurements taken in the field and by satellite.

Neven, IJIS is a geoinformatics data centre, hosted by the International Arctic Research Center (IARC) at the University of Alaska, Fairbanks (UAF). It is operated in cooperation with the Japan Aerospace Exploration Agency (JAXA), and involves Research scientists from all over the world.

Hopefully, joint International support will shelter IJIS operations against potential defunding by anti-science elements in the incoming US Congress.


Lodger, that's all fine and dandy, but I want my data. ;-)

Daniel Bailey

OK, I finally remembered...

A reader on SkS posted a question about a recent Goddard post.

I basically called it nonsense.

My question is this: Was I correct in my characterization of PIPS 2.0 as a forecasting of predicted conditions, not a snapshot of actual conditions?


The Yooper

Artful Dodger

Neven, here's another hint (can you spot the anti-science target?):

In a (March, 1997) meeting with the former Japanese Prime Minister Ryutaro Hashimoto and the former U.S. vice president Al Gore, the initiative on Global Change Research and Prediction was added to the Common Agenda for Cooperation in Global Perspective. Following the meeting, the International Arctic Research Center (IARC) was cooperatively constructed in the University of Alaska Fairbanks (UAF), under the US-Japan Science and Technology Agreement (UJST).


I don't waste time on WTFUWT. But since you've been suckered into this morass, I'll offer the following lifeline...

The mutt did his arithmetic wrong when pixel counting. He counted each pixel as 100% sea ice covered, instead of using the PIPS sea ice concentration value. This overestimates sea ice volume by about 50%, counting ice thickness where there is no ice cover. The mutt's analysis was thoroughly debunked by a NSIDC Scientist in comments on the original WTF post. Mutt then ignores this correction, saying it's okay because he's always done it this way.

Your zombie commenter on SkSci is just perpetuating a debunked denier meme. Good luck!


So, Lodger, basically what you are saying is that Al Gore invented the IJIS graph? Thank goodness, the people on the other side of the mirror finally have a good reason for discarding it completely. No need to mention the spectacular downturn we have been witnessing the past weeks! Olé! Let's focus on the Antarctic, people.


Yooper, it was fairly obvious during the summer that the PIPS 2.0 tended to interpret areas of high concentrations as high thickness on their daily plots. For example July 1st shows various patches of 5m ice along the asian shoreline that other imagers showed weren't there.


Artful Dodger

Neven, two words: 1. ESA 2. CryoSat2. De-fund that!

SIE is a specious metric for sea ice change, but it was all that could be achieved with 1979 technology. Do you still have a Floppy Disc Drive in your PC? (r u still on a PC??). Personally, I won't miss SIE. I focus on SSTs and salinity.

Have you seen how salty the Arctic Basin is getting? Arctic Salinity anomaly is a good measure of the change in sea ice age (newer ice is saltier). Vast areas of younger sea ice in the Beaufort Sea and Barents/Kara Seas. Also, notice all the Older sea ice being advected thru Fram Strait into the Greenland Sea. This older ice is a total loss as it heads south to meet it's fate in the Gulf Stream.


Lodger, at the start of the next melting season, or just before it, I will give you the following choice:

1. You write a guest blog about the whole salinity thing, what it means and how to keep an eye on it (that will save me at least 10 nights of reading, learning and writing).
2. You get banned from commenting.


Artful Dodger

Salinity is of more interest during the Winter. By Summer, I'll be watching CAPIE*. Better to ban me right now...

* The fraction of dark open water to white sea ice in Summer is a direct measure of the albedo flip which is part of Arctic Amplification. More later, but the short answer is when CAPIE goes below 70%, we are in the red zone for melt.


Hi Lodger, I'd like to see a post on salinity. I can see the lower psu on the east coast of Greenland, but what does the rest of it mean. Is salinity being generated in the water by the ice or is is flowing in from the ocean. Please enlighten us!

Artful Dodger

Hi Mike, the best way to Learn is to Teach :)

Read Chapter 6 - Temperature, Salinity, and Density from "Introduction to Physical Oceanography", an Open Source Textbook by Dr. Bob Stewart, Professor Emeritus, Texas A&M University.

You can also Search this blog, there's lots here on 'salinity' if you dig a little. Cheers!


As Professor Inferno noted in his Arctic Sea Ice Prediction, IJIS teams probably have not updated their graph because... they're choosing the new color for 2011 ! I'd bet on White ;-)
I wish us a Very Icy Arctic Year.. but I wonder it will be awful. When trying to predict its evolution with our poor models, remember that 2007 proved all them false... and that reality is very often worse than forecast on these topics.

michael sweet

On the DMI graph of temperatures north of 80N. The surface of the melting ice is 0C but the sea water under it is -2.5C. (the ice insulates the air from the water). It might be the case that when there is a lot of open water, as there was this year near the pole, the air temperature goes down because it is in contact with the colder water. With full ice coverage the temperature is higher than with partial ice coverage. Once all the ice is gone the water temperature will start to go up again.

I have not seen this discription from a real Arctic scientist, but it seems plausible to me.


Bit O/T but can anyone else get on Rapidfire? Has been down for a day or so.

L. Hamilton

Uni Bremen already has a color for this year. 2011 appears to start out with a new January minimum.


Lake Mead = Leaving Las Vegas

Dorlomin, the MODIS Arctic Mosaic seems to work for me (but there's a big white blotcvh in the middle ;-) ).


Yep got the image I was looking for from rapidfire today, it was and image of the queesland flooding.



The data is form the graphics at http://psc.apl.washington.edu/ArcticSeaiceVolume

Procedure in short:
- load bitmap in Inkscape SVG editor
- convert the trend to vector graphics
- export as XML
- import in Numeric SpreadSheet program
- Calculate monthly data points
- Convert from anomalies to absolute numbers.

If anyone wants the data, for the time being it is on http://snipt.org/wkggn


Lake Mead is recovering! -- fredt34 | January 03, 2011 at 15:28

If I look at the "little picture" it has recovered more than Four (4) Feet in level during the very short period of the month of December 2010 ( TIC = PiTR )


It may be necessary to "reload" or "refresh" browser to to get the 150 Day Blue Chart.



You have used a much more elegant approach than I did, but I believe your data has some problems. I downloaded your dataset and compared it to mine. There is pretty good agreement (within 2%) for most of the montly volumes, but I do notice some areas where we diverge quite a bit - especially in September and October, because the actual volumes are small. We pretty much agree on the September average (I make it 13406 to your 13436), but some individual years to not agree well. So you might want to check it over (or not).

Some examples where I think you have problems (figures are volume / anomaly):
In 1983, you figure volume of 15477 / anomaly of +2039, but the PIOMAS graph shows an anomaly at this time that varies between +2900 and +3300. My estimate is 16497 / +3091.
In 2004, you have 8844 / -4592, but the graph shows an anomaly just below -3000. My estimate is 10229 / -3177.
In 2006, you have 8158 / -5278, graph is between -3500 and -4500. I estimate 9672 / -3734
In 2008, you have 6848 / -6588, graph is mostly below -7000. I estimate 6126 / -7280.

I've been thinking about the technique I used to smooth out the data, and I don't think I did it optimally. If I get the chance and the enthusiasm, I might do over and post if it makes a difference. However, any differences will not negate the above examples - while I did pick them because they are the biggest discrepancies from my own dataset, checking them against the PIOMAS graph suggests that they are real discrepancies that might be reduced but not removed by me using better smoothing.

Mine here for comparison: http://snipt.org/wkuj


Andrew Xnn

FrankD and Rhk001;

Thanks for posting the numbers.
Have imported into Excel; will trend both and report back later.

What is first row of numbers labeled "0"?
Is it for 1978 or some average?

Artful Dodger

IJIS is back in operation. Look for the SIE update in about 50 minutes...

Artful Dodger

Jan 4, 2001 IJIS SIE is ... not out yet :(
Check back in 11 hours i guess.



And the winner is Red! AGAIN!

They made 2007 black and shifted the other colours forward one.

Artful Dodger

The latest value : 12,183,906 km2 (January 4, 2011)



For my data, the zero row is the monthly baseline for 1979-2009, interpolated from the mid month points here: http://psc.apl.washington.edu/ArcticSeaiceVolume/images/PIOMAS_daily_mean.png

As mentioned earlier, its far from the ideal way to do it, but it was easy and close enough for my purposes.

For Rhk001 (did I blink and miss something?), I assume its also the monthly baseline, but I wouldn't like to guess at how it was derived. Just that it was probably a better method than mine ;-)



Thanks for your data. I will be back on this.


the row "0" is the average baseline, datapoints actually from http://psc.apl.washington.edu/ArcticSeaiceVolume/images/PIOMAS_daily_mean.png


Has anyone been watching the reports from Canadian Weather Office? The temps at the moment are off the charts...

Iqaluit and Kimmirut on Baffin Island 0 to 2, 20 above average.
Sanikilauq (E Hudson Bay) to reach 0 tomorrow, around 25 above average.
Resolute to hit -3, 25 above average.
Churchill on Hudson Bay, -8, 15 above average.
Salluit (N tip of Quebec), -2, 20 above average,
Nain in Labrador is above zero, and 18 degrees above average...the list goes on.

And the overnight minima are relatively higher, some at 30 degrees above climate. Clear skies at the end of the week will see all these places plunge back to closer to normal, but there's been a few periods like this so far this winter.

Meanwhile, DMI are showing more pulses of "warm" water flowing into Baffin Bay, and Barents and Greenland Seas: http://ocean.dmi.dk/arctic/satellite/index.uk.php

Any bets on when (if?) Hudson Bay will finish freezing over?

Artful Dodger

Yes Frank, I've been watching Nunavut temps. It is +7 C now (04:00 EST Jan 5, 2011) in Pangnirtung on Baffin Island (this is 37 C above normal). Humidity is 74%. Winds are SSE 11 km/h, blowing in off of the Labrador Sea. Baffin Island usually has Northerly winds.

Even more incredible is the fact that there is 300 km of sea ice between the open ocean and the Wx station at Pangnirtung Airport. So how much heat has the air given up before it gets to Pangnirtung?

Here's what's happening. The Gulf stream normally has a meander into the Labrador Sea: (click the image to read the Wiki page)

This year the meander is large, energized by an abnormally warm gulf stream. NW Atlantic SST anomalies show how much more heat is being transported into the Labrador Sea.

Remember, the Gulf Stream carries over 60 Terawatts (TW) of heat energy into the North Atlantic. Right now, a fair swack of it is impinging on Hudson Bay, Baffin Island, and Greenland.

The result is WACCy (Warm Arctic, Cold Continents) weather. Welcome to the future!


OK fixed the bug (dates in the middle of the graphic could be 3 months off). The new data when plotted, falls now right on top of those from FrankD.

New data: http://snipt.org/wkyg

Newly generated plots:

September curve fitting:


All months in a graphic:

The new data does not change any conclusions. Note that the "exponential" curve ranks now second in the R² scores after the "logarithmic" curve. Since that curve violates physical law (infinite melt rate) the "exponential" curve is still preferred.

Lord Soth

Here is the temperture map for today Jan 5 for Canada and the Arctic

It's not just the Eastern Canadian Arctic. Most of the Arctic shoud be dark blue, however in most areas, if it is not above freezing, it's getting southern canada, northern usa winter tempertures; everywhere except the Central Canadian Arctic..


Here is an article from CBC on the arctic tempertures. It was a major headline yesterday. It talks about the lack of seaice leading to a feedback cycle.



Lodger - yes, I'm aware of WACCy weather but I'm still blown away by the impact. There's several nice diagrams doing the rounds showing the persistant high over S Greenland pumping warm air into Baffin Bay and cold air into NW Europe, and the old situation with a low doing the reverse.

For millenia, Northwestern Europe has had mild winters, considering its latitude. Superficially it is only currents and airflow that seperates the UK from the minus 20's of Quebec and Labrador. Suffice it to say it will be interesting to see whether the last couple of winters become the new "normal", are a temporary blip, or harbingers of even colder winters, such as prevail elsewhere in the 50's north.

Wipneus - good job! 99% of our data now agrees to within ~0.4% most of which comes from the differing approach. We still differ on a few single datapoints but the fact that you saw very little difference from the major ones you've already cleansed says that tweaking the data further will not change anything materially. Your curves are a bit more agressive than mine, leading to 7 monthly trends zeroing out by 2020 where only 4 of mine do. Looking at timings it would appear that you posit ice-free conditions from mid-June to quite late in December by 2020 (+/-). My curves suggest more like late July to mid-November.

One thing I find interesting - the 2010 data points are nearly all below the best fitting curve. Looked at one way, we can have a pretty good year and still be on this...well, alarming, there's no other word... trajectory. Alternatively, another year like 2010 will see those curves pulled even more steeply (ie hitting zero even earlier)

Now, I have to say again - the REAL curves are most likely S-shaped. That said, the September line is getting so close to bottoming out now, that there seems little prospect of a large tail extending the date very far. Its really only going to take another two or three years like 2010 for it to be gone. And really, when you look at the weather conditions that prevailed (as opposed to extent/area/volume/state of the ice) 2010 was a pretty soft year after June.

And all of this is predicated on PIOMAS being an accurate model.

Andrew Xnn

"And all of this is predicated on PIOMAS being an accurate model."

Good point.

How is it derived and has there been any verification of the technique?


NSIDC has a new update out.

At the end very interestingly:

NSIDC has recently released a new sea ice extent product, called the Multisensor Analyzed Sea Ice Extent (MASIE). The MASIE project is a collaborative effort with the U.S. National Ice Center to provide a daily high-resolution sea ice extent estimate based on the NOAA Interactive Mapping System (IMS) product. This new product is considered more accurate for tracking the current position of the ice edge; for climate studies, the passive microwave product (on which Arctic Sea Ice News & Analysis is based) is preferred for a more consistent, long-term average.

Pretty neat extent map! And MASIE resembles our CAPIE (compactness graph), so I like the name too!

If I've understood correctly extent numbers can from now on be downloaded every day, with extent broken up in regions as well! This could be a lot of fun (and work) to play with. But first I have to figure out how to properly import the CSV file.

Artful Dodger

"Sea ice response to an extreme negative phase of the Arctic Oscillation during winter 2009/2010"

Stroeve, Maslanik, Serreze, Rigor, & Meier
American Geophysical Union, Fall Meeting 2010


An extreme negative phase of the Arctic Oscillation (AO) characterized winter 2009/2010, leading to a strong Beaufort Gyre, a weak Transpolar Drift Stream (TDS), and near normal ice flow out of Fram Strait. In term of impacts on ice transport, autumn and winter of 2009/2010 saw above-average ice transport from north of the Canadian Archipelago westward into the eastern Beaufort and western Chukchi seas (Figure 1). Ice originating from this location is some of the oldest and thickest in the Arctic. Typically, ice that forms within, or drifts into, the Beaufort Gyre may circulate within the Canada Basin for several years, becoming thicker with time. Eventually, the thick ice enters the TDS and exits the Arctic through Fram Strait. A stronger Beaufort Gyre, as tends to accompany the negative phase of the AO, should therefore foster the development of older, thicker ice that survives summer melt, meaning more ice in September. However, a key aspect of the 2009/2010 winter transport pattern is that the winds drove the older ice directly across the Beaufort into the Chukchi Sea, as opposed to curving northward in the western Beaufort. Heading into the 2010 melt season we then had the situation where more MYI was present in the Chukchi and southern Beaufort seas than in recent years. A priori, this thicker ice should have had a better chance of surviving summer melt and thus replenished the overall MYI extent. However,by the end of August, most of the old ice transported into these regions had disappeared, further depleting the Arctic's store of old, thick ice. Ice extent in June and July set new record lows, whereas August was the second lowest ice extent during the satellite era. September ice extent ended up as the third lowest. Total areal ice transport across five gates (see inset map) in the western Beaufort (WB), eastern Beaufort (EB), western end of the region north of the Canadian Archipelago (CA) (all three positive for westward transport), Fram Strait (Fram) and Transpolar Drift Stream (TDS) (both positive for transport towards the Atlantic) for October through March. Climatology transport rates are computed from 1979 to 2009.

The presentation was delivered in San Francisco in Dec 2010. Full text is not available.


FrankD: the REAL curves are most likely S-shaped

I do not see reason for this. Especially for the September/August curves. The rate of ice shrinkage is significantly increasing, most likely from some self amplifying mechanism. That does not stop until the ice is gone.
After the late summer ice has disappeared, the self amplification may slow down, so the ice in other parts of the year could stay longer than we now predict. For those months you could be right.

the 2010 data points are nearly all below the best fitting curve
I would like to add another point that you may want to consider. The self amplifying mechanism that I proposed for the "exponential" curve must have kicked in at some recent time. If not, it would be hard to explain the more or less stable Arctic ice cover in the last hundreds or even thousands of years.
This is also what you see: restricting data points from 1996 to present will raise R2 to 0.96 indicating a better fit. I am in no position to declare this statistically significant, but I did use in my estimate for ice free September conditions in 2014 (range 2013-2016).

The curves for dates after 1996 are shown in http://img228.imageshack.us/img228/2294/piomastrnd2.png .
As you see, if the previous curves are alarming these become really dramatic, in 2020 only a bit of ice in April.

Again, if rates in the cold part of the year slow down when summer ice is gone, the ice may linger a bit longer than that.

Of course, I agree that this all depends on the fidelity of the PIOMAS estimates.

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.


Post a comment